AT&T uses Nvidia, Google and Meta open models for AI costs

AT&T said it is using open AI models from Nvidia, Google and Meta to keep its AI costs steady as usage increases.

TL;DR

• AT&T said open AI models from Nvidia, Google and Meta are part of its AI cost strategy.
• The operator described the approach as managing AI “tokenomics,” or the cost of running AI workloads as usage grows.
• The report came from Fierce Network and did not specify deployment dates, spending figures, or named products.

AT&T is using open models from U.S. companies Nvidia, Google and Meta as part of what it calls its AI “tokenomics” strategy. In this context, tokenomics refers to managing the cost of AI usage, including the tokens processed by large language models as demand rises.

The stated goal is to keep costs steady while AT&T expands its use of AI. The source description did not provide figures for token volumes, infrastructure spending, model sizes, or the business units using the models.

The available source material identifies Nvidia, Google and Meta as the model providers and AT&T as the operator adopting them. The source material did not state whether AT&T is using proprietary models alongside open models, nor did it name any cloud provider, deployment timeline, or customer-facing service.

Related questions

• Which open AI model providers did AT&T name in its cost strategy?
• What does AT&T mean by AI “tokenomics”?
• Did the source provide spending figures or deployment dates for AT&T’s AI models?